The conventional network architecture has a hierarchical structure of two or more layers. As shown in FIG. 1, a network can be divided into different layers (including a core layer, a convergence layer, an edge access layer, and a peer layer) according to a function and a deployment location of each layer, the reliability between layers is generally improved by adopting the dual-homing redundant connection, and traffics of users are converged layer by layer. The access layer is responsible for network access of a user terminal and used for providing abundant user interface types, and has a wide node distribution and a high interface density. The convergence layer is responsible for converging traffics of access nodes and used for expanding service coverage range of a core node, and has abundant interface types, a strong convergence capability, and an integrated service processing capability. The core layer is responsible for high-speed forwarding and inter-area service interworking, and has a small number of nodes.
In order to integrate and simplify the network structure and improve resource utilization, in the existing network architecture, when a point-to-multipoint switching packet is processed, a method is used in which two or more master switching devices in the same network layer form a virtual network device to reduce the complexity caused by using multiple redundant components.
In addition, in an existing hardware clustered router, in order to adapt to a huge data exchange capacity between devices in the cluster, multiple high-end routers are connected to a central switching matrix frame chassis through an ultra-short-range optical fiber, to form a large clustered router of one fabric card chassis with four linecard chassis, two fabric card chassis with eight linecard chassis, or four fabric card chassis with sixteen linecard chassis. Due to the introduction of the dedicated central switching matrix frame and the very short reach optical fiber interconnection technology, the network topology structure is complicated, and the routing convergence and stability are poor.
To sum up, the processing for a multicast packet is not so sound at present, regardless of the virtual cluster convergence of the devices on the same network layer for simplifying the network structure, or the hardware clustered router applied at the core node or a super node.